The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to run npx vigiles compile and npx vigiles lint to process and verify specification files.\n- [EXTERNAL_DOWNLOADS]: The use of npx involves the npm registry, which may download the vigiles package at runtime.\n- [REMOTE_CODE_EXECUTION]: Executing code via npx is a form of remote code execution, as it fetches and runs packages from an external registry.\n- [INDIRECT_PROMPT_INJECTION]: User input provided via $ARGUMENTS is incorporated into .spec.ts files, which are compiled into the core instruction files CLAUDE.md and AGENTS.md. This path allows untrusted input to influence future agent behavior.\n
Ingestion points: User input passed via skill arguments in SKILL.md.\n
Boundary markers: None identified in the instruction generation logic to separate user-contributed rules from system-level instructions.\n
Capability inventory: The skill provides access to file manipulation tools (Read, Edit, Write) and shell execution (Bash).\n
Sanitization: No explicit validation or escaping of user-provided content is performed before adding it to the project specification.

edit-spec